Magnetic Resonance Image Quality Research Articles

Continuum topological derivative - a novel application tool for denoising CT and MRI medical images

BackgroundCT and MRI modalities are important diagnostics tools for exploring the anatomical and tissue properties, respectively of the human beings. Several advancements like HRCT, FLAIR and Propeller have advantages in diagnosing the diseases very accurately, but still have enough space for improvements due to the presence of inherent and instrument noises. In the case of CT and MRI, the quantum mottle and the Gaussian and Rayleigh noises, respectively are still present in their advanced modalities of imaging. This paper addresses the denoising problem with continuum topological derivative technique and proved its trustworthiness based on the comparative study with other traditional filtration methods such as spatial, adaptive, frequency and transformation techniques using measures like visual inspection and performance metrics.MethodsThis research study focuses on identifying a novel method for denoising by testing different filters on HRCT (High-Resolution Computed Tomography) and MR (Magnetic Resonance) images. The images were acquired from the Image Art Radiological Scan Centre using the SOMATOM CT and SIGNA Explorer (operating at 1.5 Tesla) machines. To compare the performance of the proposed CTD (Continuum Topological Derivative) method, various filters were tested on both HRCT and MR images. The filters tested for comparison were Gaussian (2D convolution operator), Wiener (deconvolution operator), Laplacian and Laplacian diagonal (2nd order partial differential operator), Average, Minimum, and Median (ordinary spatial operators), PMAD (Anisotropic diffusion operator), Kuan (statistical operator), Frost (exponential convolution operator), and HAAR Wavelet (time–frequency operator). The purpose of the study was to evaluate the effectiveness of the CTD method in removing noise compared to the other filters. The performance metrics were analyzed to assess the diligence of noise removal achieved by the CTD method. The primary outcome of the study was the removal of quantum mottle noise in HRCT images, while the secondary outcome focused on removing Gaussian (foreground) and Rayleigh (background) noise in MR images. The study aimed to observe the dynamics of noise removal by examining the values of the performance metrics.In summary, this study aimed to assess the denoising ability of various filters in HRCT and MR images, with the CTD method being the proposed approach. The study evaluated the performance of each filter using specific metrics and compared the results to determine the effectiveness of the CTD method in removing noise from the images.ResultsBased on the calculated performance metric values, it has been observed that the CTD method successfully removed quantum mottle noise in HRCT images and Gaussian as well as Rayleigh noise in MRI. This can be evidenced by the PSNR (Peak Signal-to-Noise Ratio) metric, which consistently exhibited values ranging from 50 to 65 for all the tested images. Additionally, the CTD method demonstrated remarkably low residual values, typically on the order of e−09, which is a distinctive characteristic across all the images. Furthermore, the performance metrics of the CTD method consistently outperformed those of the other tested methods. Consequently, the results of this study have significant implications for the quality, structural similarity, and contrast of HRCT and MR images, enabling clinicians to obtain finer details for diagnostic purposes.ConclusionContinuum topological derivative algorithm is found to be constructive in removing prominent noises in both CT and MRI images and can serve as a potential tool for recognition of anatomical details in case of diseased and normal ones. The results obtained from this research work are highly inspiring and offer great promise in obtaining accurate diagnostic information for critical cases such as Thoracic Cavity Carina, Brain SPI Globe Lens 4th Ventricle, Brain-Middle Cerebral Artery, Brain-Middle Cerebral Artery and neoplastic lesions. These findings lay the foundation for implementing the proposed CTD technique in routine clinical diagnosis.

Impact of prostate MRI image quality on diagnostic performance for clinically significant prostate cancer (csPCa).

With the widespread clinical application of prostate magnetic resonance imaging (MRI), there has been an increasing demand for lesion detection and accurate diagnosis in prostate MR, which relies heavily on satisfactory image quality. Focusing on the primary sequences involved in Prostate Imaging Reporting and Data System (PI-RADS), this study have evaluated common quality issues in clinical practice (such as signal-to-noise ratio (SNR), artifacts, boundaries, and enhancement). The aim of the study was to determine the impact of image quality on clinically significant prostate cancer (csPCa) detection, positive predictive value (PPV) and radiologist's diagnosis in different sequences and prostate zones. This retrospective study included 306 patients who underwent prostate MRI with definitive pathological reports from February 2021 to December 2022. All histopathological specimens were evaluated according to the recommendations of the International Society of Urological Pathology (ISUP). An ISUP Grade Group ≥ 2 was considered as csPCa. Three radiologists from different centers respectively performed a binary classification assessment of image quality in the following ten aspects: (1) T2WI in the axial plane: SNR, prostate boundary conditions, the presence of artifacts; (2) T2WI in the sagittal or coronal plane: prostate boundary conditions; (3) DWI: SNR, delineation between the peripheral and transition zone, the presence of artifacts, the matching of DWI and T2WI images; (4) DCE: the evaluation of obturator artery enhancement, the evaluation of dynamic contrast enhancement. Fleiss' Kappa was used to determine the inter-reader agreement. Wilson's 95% confidence interval (95% CI) was used to calculate PPV. Chi-square test was used to calculate statistical significance. A p-value < 0.05 was considered statistically significant. High-quality images had a higher csPCa detection rate (56.5% to 64.3%) in axial T2WI, DWI, and DCE, with significant statistical differences in SNR in axial T2WI (p 0.002), the presence of artifacts in axial T2WI (p 0.044), the presence of artifacts in DWI (p < 0.001), and the matching of DWI and T2WI images (p < 0.001). High-quality images had a higher PPV (72.5% to 78.8%) and showed significant statistical significance in axial T2WI, DWI, and DCE. Additionally, we found that PI-RADS 3 (24.0% to 52.9%) contained more low-quality images compared to PI-RADS 4-5 (20.6% to 39.3%), with significant statistical differences in the prostate boundary conditions in axial T2WI (p 0.048) and the presence of artifacts in DWI (p 0.001). Regarding the relationship between csPCa detection and image quality in different prostate zones, this study found that significant statistical differences were only observed between high- (63.5% to 75.7%) and low-quality (30.0% to 50.0%) images in the peripheral zone (PZ). Prostate MRI quality may have an impact on the diagnostic performance. The poorer image quality is associated with lower csPCa detection rates and PPV, which can lead to an increase in radiologist's ambiguous diagnosis (PI-RADS 3), especially for the lesions located at PZ.

Inter-scanner super-resolution of 3D cine MRI using a transfer-learning network for MRgRT

Objective. Deep-learning networks for super-resolution (SR) reconstruction enhance the spatial-resolution of 3D magnetic resonance imaging (MRI) for MR-guided radiotherapy (MRgRT). However, variations between MRI scanners and patients impact the quality of SR for real-time 3D low-resolution (LR) cine MRI. In this study, we present a personalized super-resolution (psSR) network that incorporates transfer-learning to overcome the challenges in inter-scanner SR of 3D cine MRI. Approach: Development of the proposed psSR network comprises two-stages: (1) a cohort-specific SR (csSR) network using clinical patient datasets, and (2) a psSR network using transfer-learning to target datasets. The csSR network was developed by training on breath-hold and respiratory-gated high-resolution (HR) 3D MRIs and their k-space down-sampled LR MRIs from 53 thoracoabdominal patients scanned at 1.5 T. The psSR network was developed through transfer-learning to retrain the csSR network using a single breath-hold HR MRI and a corresponding 3D cine MRI from 5 healthy volunteers scanned at 0.55 T. Image quality was evaluated using the peak-signal-noise-ratio (PSNR) and the structure-similarity-index-measure (SSIM). The clinical feasibility was assessed by liver contouring on the psSR MRI using an auto-segmentation network and quantified using the dice-similarity-coefficient (DSC). Results. Mean PSNR and SSIM values of psSR MRIs were increased by 57.2% (13.8–21.7) and 94.7% (0.38–0.74) compared to cine MRIs, with the reference 0.55 T breath-hold HR MRI. In the contour evaluation, DSC was increased by 15% (0.79–0.91). Average time consumed for transfer-learning was 90 s, psSR was 4.51 ms per volume, and auto-segmentation was 210 ms, respectively. Significance. The proposed psSR reconstruction substantially increased image and segmentation quality of cine MRI in an average of 215 ms across the scanners and patients with less than 2 min of prerequisite transfer-learning. This approach would be effective in overcoming cohort- and scanner-dependency of deep-learning for MRgRT.

3D‐SRDM: 3D super‐resolution of MRI volumes based on diffusion model

AbstractBrain magnetic resonance imaging (MRI) is crucial for diagnosing and understanding neurological disorders, but inherent limitations hinder the visualization of fine details in brain structures. The emergence of super‐resolution techniques, especially deep‐learning methods, has improved imaging quality of MRI, by increasing MRI spatial resolution. At present, deep‐learning algorithms mostly performed super‐resolution on 2D MRI images. However, considering 3D nature of MRI, 3D models are more suitable for brain MRI super‐resolution. To achieve finer brain structural details, this study proposes a 3D brain MRI super‐resolution method based on diffusion model (3D‐SRDM), which is a fast and easily trainable neural network for the generation of high‐resolution brain MRI images. In our 3D‐SRDM model, the self‐attention module in U‐Net is replaced with 3D spatial attention mechanism. The network structure of 3D‐SRDM is optimized to reduce training parameters. Moreover, accelerated sampling from denoising diffusion implicit model is also incorporated to reduce time consumption. By these optimizations, compared with original diffusion model, the proposed model can achieve about 10‐ and 5‐fold speed increase at 4× and 8× super‐resolution of 3D brain MRI volumes, respectively, almost without affecting image quality. Thus, 3D‐SRDM has potential application value in efficiently generating high‐resolution 3D brain MRI images, thus facilitating the doctors' diagnosis.

Toward an optimal cadaveric brain model for neurosurgical education: assessment of preservation, parenchyma, vascular injection, and imaging.

We assessed types of cadaveric head and brain tissue specimen preparations that are used in a high throughput neurosurgical research laboratory to determine optimal preparation methods for neurosurgical anatomical research, education, and training. Cadaveric specimens (N = 112) prepared using different preservation and vascular injection methods were imaged, dissected, and graded by 11 neurosurgeons using a 21-point scale. We assessed the quality of tissue and preservation in both the anterior and posterior circulations. Tissue quality was evaluated using a 9-point magnetic resonance imaging (MRI) scale. Formalin-fixed specimens yielded the highest scores for assessment (mean ± SD [17.0 ± 2.8]) vs. formalin-flushed (17.0 ± 3.6) and MRI (6.9 ± 2.0). Cadaver assessment and MRI scores were positively correlated (P < 0.001, R2 0.60). Analysis showed significant associations between cadaver assessment scores and specific variables: nonformalin fixation (β = -3.3), preservation within ≤72h of death (β = 1.8), and MRI quality score (β = 0.7). Formalin-fixed specimens exhibited greater hardness than formalin-flushed and nonformalin-fixed specimens (P ≤ 0.006). Neurosurgeons preferred formalin-flushed specimens injected with colored latex. For better-quality specimens for neurosurgical education and training, formalin preservation within ≤72h of death was preferable, as was injection with colored latex. Formalin-flushed specimens more closely resembled live brain parenchyma. Assessment scores were lower for preparation techniques performed > 72h postmortem and for nonformalin preservation solutions. The positive correlation between cadaver assessment scores and our novel MRI score indicates that donation organizations and institutional buyers should incorporate MRI as a screening tool for the selection of high-quality specimens.

Effect of Breath Training on Image Quality of Chest Magnetic Resonance Free-breathing Sequence.

Magnetic Resonance Imaging (MRI) plays a role in demonstrating substantial utility in lung lesion imaging, detection, diagnosis, and evaluation. Previous studies have found that free-breathing star VIBE sequences not only have high image quality but also have a high ability to detect and display nodules. However, in our routine clinical practice, we have encountered suboptimal image quality in the free-breathing sequences of certain patients. This study aims to assess the impact of breath training on the quality of chest magnetic resonance imaging obtained during free-breathing sequences. A total of 68 patients with lung lesions, such as nodules or masses detected via Computed Tomography (CT) examination, were prospectively gathered. They were then randomly divided into two groups: an observation group and a control group. Standard preparation was performed for all patients in both groups before the examination. The observation group underwent 30 minutes of breath training prior to the MRI examination additionally, followed by the acquisition of MRI free-breathing sequence images. The signal intensity (SI) and standard deviation (SD) of the lesion and adjacent normal lung tissue were measured, and the image signal-to-noise ratio (SNR) and contrast signal-to-noise ratio (CNR) of the lesion were calculated for objective image quality evaluation. The subjective image quality of the two groups of images was also evaluated using a 5-point method. MRI examinations were completed in both groups. Significantly better subjective image quality (edge and internal structure clarity, vascular clarity, breathing and cardiac artifacts, and overall image quality) was achieved in the observation group compared to the control group (P<0.05). In addition, higher SNR and CNR values for disease lesions were observed in the observation group compared to the control group (t=4.35, P<0.05; t=5.35, P<0.05). It is concluded that the image quality of free-breathing sequences MRI can be improved through breath training before examination.

Prostate MRI for the detection of clinically significant prostate cancer: Update and future directions.

In recent decades, there has been an increasing role for magnetic resonance imaging (MRI) in the detection of clinically significant prostate cancer (csPC). The purpose of this review is to provide an update and outline future directions for the role of MRI in the detection of csPC. In diagnosing clinically significant prostate cancer pre-biopsy, advances include our understanding of MRI-targeted biopsy, the role of biparametric MRI (non-contrast) and changing indications, for example the role of MRI in screening for prostate cancer. Furthermore, the role of MRI in identifying csPC is maturing, with emphasis on standardization of MRI reporting in active surveillance (PRECISE), clinical staging (EPE grading, MET-RADS-P) and recurrent disease (PI-RR, PI-FAB). Future directions of prostate MRI in detecting csPC include quality improvement, artificial intelligence and radiomics, positron emission tomography (PET)/MRI and MRI-directed therapy. The utility of MRI in detecting csPC has been demonstrated in many clinical scenarios, initially from simply diagnosing csPC pre-biopsy, now to screening, active surveillance, clinical staging, and detection of recurrent disease. Continued efforts should be undertaken not only to emphasize the reporting of prostate MRI quality, but to standardize reporting according to the appropriate clinical setting.

MR-Safe Cartesian Platform for Active Cardiac Shimming: Preliminary Validation.

Implanted Cardioverter Defibrillators (ICDs) induce a large (100 parts per million) inhomogeneous magnetic field in the magnetic resonance imaging (MRI) scanner which cannot be corrected by the scanner's built-in shim coils, leading to significant image artifacts that can make portions of the heart unreadable. To compensate for the field inhomogeneity, an active shim coil capable of countering the field deviation in user-defined regions was designed that must be optimally placed at patient-specific locations. We aim to develop and evaluate an MR-safe robotic solution for automated shim coil positioning. We designed and fabricated an MR-safe Cartesian platform that holds the shim coil inside the scanner. The platform consists of three lead screw stages actuated by pneumatic motors, achieving decoupled translations of 140 mm in each direction. The platform is made of plastics and fiberglass with the control electronics placed outside the scanner room, ensuring MR safety. Mechanical modeling was derived to provide design specifications. Experiments show that the platform achieves less than 2 mm average motion error and 0.5 mm repeatability in all directions, and reduces the adjustment time from 5 min to a few seconds. Phantom and animal trials were conducted, showing that the proposed system is able to position a heavy shim coil ( kg) for improved ICD artifact suppression. This robotic platform provides an effective method for reliable shim coil positioning inside the scanner. This work contributes to improving cardiac MRI quality that could facilitate accurate diagnosis and treatment planning for patients with implanted ICDs.

Endograft position and endoleak detection after endovascular abdominal aortic repair with low-field tiltable MRI: a feasibility study

BackgroundAbdominal aortic endoleaks after endovascular aneurysm repair might be position-dependent, therefore undetectable using supine imaging. We aimed to determine the feasibility and benefit of using a low-field tiltable magnetic resonance imaging (MRI) scanner allowing to study patients who can be imaged in both supine and upright positions of endoleaks.MethodsTen EVAR patients suspected of endoleak based on ultrasound examination were prospectively included. MRI in upright and supine positions was compared with routine supine computed tomography angiography (CTA). Analysis was performed through (1) subjective image quality assessment by three observers, (2) landmark registration between MRI and CTA scans, (3) Euclidean distances between renal and endograft landmarks, and (4) evaluation of endoleak detection on MRI by a consensus panel. Statistical analysis was performed by one-way repeated measures analysis of variance.ResultsThe image quality of upright/supine MRI was inferior compared to CTA. Median differences in both renal and endograft landmarks were approximately 6–7 mm between upright and supine MRI and 5–6 mm between supine MRI and CTA. In the proximal sealing zone of the endograft, no differences were found among all three scan types (p = 0.264). Endoleak detection showed agreement between MRI and CTA in 50% of the cases, with potential added value in only one patient.ConclusionsThe benefit of low-field upright MRI for endoleak detection was limited. While MRI assessment was non-inferior to standard CTA in detecting endoleaks in selected cases, improved hardware and sequences are needed to explore the potential of upright MRI in patients with endoleaks.Relevance statementUpright low-field MRI has limited clinical value in detecting position-dependent endoleaks; improvements are required to fulfil its potential as a complementary modality in this clinical setting.Key points• Upright MRI shows potential for imaging endoleaks in aortic aneurysm patients in different positions.• The image quality of upright MRI is inferior to current techniques.• Upright MRI complements CTA, but lacks accurate deformation measurements for clinical use.• Advancements in hardware and imaging sequences are needed to fully utilise upright MRI capabilities.Graphical

Improvement of 2D cine image quality using 3D priors and cycle generative adversarial network for low field MRI-guided radiation therapy.

Cine magnetic resonance (MR) images have been used for real-time MR guided radiation therapy (MRgRT). However, the onboard MR systems with low-field strength face the problem of limited image quality. To improve the quality of cine MR images in MRgRT using prior image information provided by the patient planning and positioning MR images. This study employed MR images from 18 pancreatic cancer patients who received MR-guided stereotactic body radiation therapy. Planning 3D MR images were acquired during the patient simulation, and positioning 3D MR images and 2D sagittal cine MR images were acquired before and during the beam delivery, respectively. A deep learning-based framework consisting of two cycle generative adversarial networks (CycleGAN), Denoising CycleGAN and Enhancement CycleGAN, was developed to establish the mapping between the 3D and 2D MR images. The Denoising CycleGAN was trained to first denoise the cine images using the time domain cine image series, and the Enhancement CycleGAN was trained to enhance the spatial resolution and contrast by taking advantage of the prior image information from the planning and positioning images. The denoising performance was assessed by signal-to-noise ratio (SNR), structural similarity index measure, peak SNR, blind/reference-less image spatial quality evaluator (BRISQUE), natural image quality evaluator, and perception-based image quality evaluator scores. The quality enhancement performance was assessed by the BRISQUE and physician visual scores. In addition, the target contouring was evaluated on the original and processed images. Significant differences were found for all evaluation metrics after Denoising CycleGAN processing. The BRISQUE and visual scores were also significantly improved after sequential Denoising and Enhancement CycleGAN processing. In target contouring evaluation, Dice similarity coefficient, centroid distance, Hausdorff distance, and average surface distance values were significantly improved on the enhanced images. The whole processing time was within 20ms for a typical input image size of 512 × 512. Taking advantage of the prior high-quality positioning and planning MR images, the deep learning-based framework enhanced the cine MR image quality significantly, leading to improved accuracy in automatic target contouring. With the merits of both high computational efficiency and considerable image quality enhancement, the proposed method may hold important clinical implication for real-time MRgRT.

Deep Learning Approaches for Automatic Quality Assurance of Magnetic Resonance Images Using ACR Phantom

BackgroundIn recent years, there has been a growing trend towards utilizing Artificial Intelligence (AI) and machine learning techniques in medical imaging, including for the purpose of automating quality assurance. In this research, we aimed to develop and evaluate various deep learning-based approaches for automatic quality assurance of Magnetic Resonance (MR) images using the American College of Radiology (ACR) standards.MethodsThe study involved the development, optimization, and testing of custom convolutional neural network (CNN) models. Additionally, popular pre-trained models such as VGG16, VGG19, ResNet50, InceptionV3, EfficientNetB0, and EfficientNetB5 were trained and tested. The use of pre-trained models, particularly those trained on the ImageNet dataset, for transfer learning was also explored. Two-class classification models were employed for assessing spatial resolution and geometric distortion, while an approach classifying the image into 10 classes representing the number of visible spokes was used for the low contrast.ResultsOur results showed that deep learning-based methods can be effectively used for MR image quality assurance and can improve the performance of these models. The low contrast test was one of the most challenging tests within the ACR phantom.ConclusionsOverall, for geometric distortion and spatial resolution, all of the deep learning models tested produced prediction accuracy of 80% or higher. The study also revealed that training the models from scratch performed slightly better compared to transfer learning. For the low contrast, our investigation emphasized the adaptability and potential of deep learning models. The custom CNN models excelled in predicting the number of visible spokes, achieving commendable accuracy, recall, precision, and F1 scores.

The impact of quality control on cortical morphometry comparisons in autism.

Structural magnetic resonance imaging (MRI) quality is known to impact and bias neuroanatomical estimates and downstream analysis, including case-control comparisons, and a growing body of work has demonstrated the importance of careful quality control (QC) and evaluated the impact of image and image-processing quality. However, the growing size of typical neuroimaging datasets presents an additional challenge to QC, which is typically extremely time and labour intensive. One of the most important aspects of MRI quality is the accuracy of processed outputs, which have been shown to impact estimated neurodevelopmental trajectories. Here, we evaluate whether the quality of surface reconstructions by FreeSurfer (one of the most widely used MRI processing pipelines) interacts with clinical and demographic factors. We present a tool, FSQC, that enables quick and efficient yet thorough assessment of outputs of the FreeSurfer processing pipeline. We validate our method against other existing QC metrics, including the automated FreeSurfer Euler number, two other manual ratings of raw image quality, and two popular automated QC methods. We show strikingly similar spatial patterns in the relationship between each QC measure and cortical thickness; relationships for cortical volume and surface area are largely consistent across metrics, though with some notable differences. We next demonstrate that thresholding by QC score attenuates but does not eliminate the impact of quality on cortical estimates. Finally, we explore different ways of controlling for quality when examining differences between autistic individuals and neurotypical controls in the Autism Brain Imaging Data Exchange (ABIDE) dataset, demonstrating that inadequate control for quality can alter results of case-control comparisons.

Independent external validation of a stroke recurrence score in patients with embolic stroke of undetermined source

BackgroundEmbolic stroke of undetermined source (ESUS) accounts for a substantial proportion of ischaemic strokes. A stroke recurrence score has been shown to predict the risk of recurrent stroke in patients with ESUS based on a combination of clinical and imaging features. This study aimed to externally validate the performance of the ESUS recurrence score using data from a randomized controlled trial.MethodsThe validation dataset consisted of eligible stroke patients with available magnetic resonance imaging (MRI) data enrolled in the PreDAFIS sub-study of the MonDAFIS study. The score was calculated using three variables: age (1 point per decade after 35 years), presence of white matter hyperintensities (2 points), and multiterritorial ischaemic stroke (3 points). Patients were assigned to risk groups as described in the original publication. The model was evaluated using standard discrimination and calibration methods.ResultsOf the 1054 patients, 241 (22.9%) were classified as ESUS. Owing to insufficient MRI quality, three patients were excluded, leaving 238 patients (median age 65.5 years [IQR 20.75], 39% female) for analysis. Of these, 30 (13%) patients experienced recurrent ischaemic stroke or transient ischemic attack (TIA) during a follow-up period of 383 patient-years, corresponding to an incidence rate of 7.8 per 100 patient-years (95% CI 5.3–11.2). Patients with an ESUS recurrence score value of ≥ 7 had a 2.46 (hazard ratio (HR), 95% CI 1.02–5.93) times higher risk of stroke recurrence than patients with a score of 0–4. The cumulative probability of stroke recurrence in the low-(0–4), intermediate-(5–6), and high-risk group (≥ 7) was 9%, 13%, and 23%, respectively (log-rank test, χ2 = 4.2, p = 0.1).ConclusionsThis external validation of a published scoring system supports a threshold of ≥ 7 for identifying ESUS patients at high-risk of stroke recurrence. However, further adjustments may be required to improve the model’s performance in independent cohorts. The use of risk scores may be helpful in guiding extended diagnostics and further trials on secondary prevention in patients with ESUS.Trial registration: Clinical Trials, NCT02204267. Registered 30 July 2014, https://clinicaltrials.gov/ct2/show/NCT02204267.

Non-sedated fast spine magnetic resonance imaging in pediatric patients.

Traditional spine magnetic resonance imaging (MRI) protocols require sedation in young children and uncooperative patients. There is an increased interest in non-sedated pediatric MRI protocols to reduce risks associated with anesthetic agents and improve MRI access. To evaluate the image quality of pediatric non-sedated fast spine MRI. We retrospectively reviewed 69 pediatric non-sedated fast spine MRI exams performed in 57 patients. Two blinded readers provided image quality ratings for the evaluation of bones, cranio-cervical junction, cerebrospinal fluid (CSF) spaces, spinal cord, soft tissues, ligaments, and overall diagnostic quality on a 1-5 scale, and determined whether there was evidence of syringomyelia, abnormal conus medullaris position, or filum terminale abnormality. Mean patient age was 7.2years (age range ≤ 1-17). Indications included syringomyelia (n=25), spinal dysraphism (n=4), combination of both syringomyelia and spinal dysraphism (n=8), and other miscellaneous indications (n=32). The inter-observer agreement ranged between moderate and very good for each variable (Cohen's weighted kappa] range=0.45-0.69). The highest image quality ratings were given to CSF spaces (mean image quality=3.5/5 ± 0.8) and cranio-cervical junction evaluations (3.5/5 ± 0.9). Overall diagnostic quality was worst in the <5years group (P=0.006). Readers independently identified a cervical spinal cord syrinx in 6 cases, and 1mm spinal cord central canal dilation in one case. Readers agreed on the position of the conus medullaris in 92% of cases (23/25 cases). Non-sedated pediatric spine MRI can be an effective diagnostic test to evaluate for spine pathology, especially syringomyelia, Chiari malformation, and conus medullaris anatomy.

Evaluation of Compressed SENSE on Image Quality and Reduction of MRI Acquisition Time: A Clinical Validation Study

To evaluate the effect of compressed SENSE (CS)in clinical settings on scan time reduction and image quality. Ninety-five magnetic resonance imaging (MRI) scans from different anatomical regions were acquired, consisting of a standard protocol sequence (SS) and sequence accelerated with CS. Anonymized paired sequences were randomly displayed and rated by six blinded subspecialty radiologists. Side-by-side evaluation on perceived sharpness, perceived signal-to-noise-ratio (SNR), lesion conspicuity, and artifacts were compared and scored on a five-point Likert scale, and individual image quality was evaluated on a four-point Likert scale. CS reduced overall scan time by 32% while maintaining acceptable MRI quality for all regions. The largest time savings were seen in the spine (mean = 68 seconds, 44% reduction) followed by the brain (mean = 86seconds, 37% reduction). The sequence with maximum time savings was intracranial 3D-time-of-flight magnetic resonance angiography (202seconds, 56% reduction). CS was mildly inferior to SS on perceived sharpness, perceived SNR, and lesion conspicuity (mean scores = 2.32-2.96, P<.001 [1:SS superior; 3: equivalent; 5: CS superior]). CS was equivalent to SS for joint and body scans on overall image quality (CS = 3.02-3.37, SS = 3.04-3.68, P>.05, [1: lowest quality and 4: highest quality]). The overall image quality of CS was slightly less for brain and spine scans (mean CS = 2.79-3.05, mean SS = 3.13-3.43, P=.021) but still diagnostic. Good overall clinical acceptance for CS (88%) was noted with full clinical acceptance for body scans (100%)and high acceptance for other regions (68%-95%). CS significantly reduced MR acquisition time while maintaining acceptable image quality. The implementation of CS may improve departmental workflows and enhance patient care.

Automatic brain MRI motion artifact detection based on end-to-end deep learning is similarly effective as traditional machine learning trained on image quality metrics.

Head motion artifacts in magnetic resonance imaging (MRI) are an important confounding factor concerning brain research as well as clinical practice. For this reason, several machine learning-based methods have been developed for the automatic quality control of structural MRI scans. Deep learning offers a promising solution to this problem, however, given its data-hungry nature and the scarcity of expert-annotated datasets, its advantage over traditional machine learning methods in identifying motion-corrupted brain scans is yet to be determined. In the present study, we investigated the relative advantage of the two methods in structural MRI quality control. To this end, we collected publicly available T1-weighted images and scanned subjects in our own lab under conventional and active head motion conditions. The quality of the images was rated by a team of radiologists from the point of view of clinical diagnostic use. We present a relatively simple, lightweight 3D convolutional neural network trained in an end-to-end manner that achieved a test set (N=411) balanced accuracy of 94.41% in classifying brain scans into clinically usable or unusable categories. A support vector machine trained on image quality metrics achieved a balanced accuracy of 88.44% on the same test set. Statistical comparison of the two models yielded no significant difference in terms of confusion matrices, error rates, or receiver operating characteristic curves. Our results suggest that these machine learning methods are similarly effective in identifying severe motion artifacts in brain MRI scans, and underline the efficacy of end-to-end deep learning-based systems in brain MRI quality control, allowing the rapid evaluation of diagnostic utility without the need for elaborate image pre-processing.

Poster 264: Medial Tibia Bone Marrow Edema on Preoperative MRI Predicts Inferior Patient Outcomes Following Posterior Medial Meniscus Root Repair

Objectives: Repair of posterior medial meniscus root (PMMR) tears demonstrates favorable patient outcomes and may prevent rapid progression of knee osteoarthritis, however, there is a paucity of data in literature regarding prognostic factors affecting postoperative outcomes. The medial meniscus plays an important role in the knee biomechanically by increasing tibiofemoral contact area and decreasing peak tibiofemoral contact pressures. Few studies exist that examine preoperative knee magnetic resonance imaging (MRI) findings and patient reported outcome measures (PROMs) following PMMR repair. Semi-quantitative MRI evaluation can offer improved detection of degenerative pathology. Identifying specific factors that may predict surgical outcomes is important when considering PMMR repair, as some patients with more advanced degenerative joint disease may benefit from consideration for knee arthroplasty instead of joint preservation with meniscus repair surgery. The purpose of this study was to identify factors on preoperative MRI that may predict postoperative outcomes following PMMR repair. We hypothesize that patients with worse knee overall quality would have worse postoperative PRO scores. Methods: Between 2012 and 2020, patients who underwent posterior medial meniscus root repair at a tertiary referral academic center by fellowship trained sports medicine surgeons were retrospectively identified, recruited, and enrolled in the study. Institutional review board approval and informed consent was obtained. Inclusion criteria included patients ages 18-70, posterior medial meniscus root repair through transtibial fixation technique, minimum two years postoperative follow-up, and a preoperative knee MRI. Exclusion criteria included prior ipsilateral knee meniscus or ligamentous injury, additional ipsilateral knee injury requiring surgery, and conversion to knee arthroplasty. Visual Analog Scale (VAS) pain and PRO surveys including Patient Reported Outcome Measures Information System Physical Function (PROMIS-PF) Computer Adaptive Test, Lysholm Knee Score, Knee Osteoarthritis and Injury Outcome Score (KOOS) with 5 sub-sections, were collected at minimum 2-years postoperative. Patient Acceptable State Score (PASS) values from a prior meniscus study were utilized to identify patients that met PASS for the KOOS subsections. A fellowship-trained musculoskeletal radiologist reviewed preoperative MRIs and calculated Whole Organ Magnetic Resonance Imaging Scores (WORMS) for meniscus, cartilage, bone marrow edema pattern (BMEP) in 6 knee locations as well as effusion/synovitis and meniscal extrusion. Radiographs were scored with Kellgren-Lawrence (KL) scores to assess presence and severity of osteoarthritis. Statistical analysis was performed with Stata (StataCorp) using two sample T-tests, Mann-Whitney, and Fisher’s exact test for categorical variables. Significance was defined as p&lt;0.05. Results: We evaluated 29 knees (22 female patients) with a mean age at surgery of 52.3±9.9 years, body mass index of 27.6±5.6 kg/m2, and at a mean follow-up of 59.6 months ± 26.5. median KL grade 2, and 44.8% with meniscal extrusion. Pain scores decreased significantly from before surgery to final follow-up (p&lt;0.001), and the percentage of patients meeting PASS ranged from 44.8% for KOOS-Sport subscale to 72.4% for KOOS-Pain and KOOS-Quality of Life subscales (Table 1). There were 10 patients with medial tibial BMEP on pre-operative MRI and 19 with no medial tibial BMEP (Table 2). Patients with medial tibia BMEP had significantly lower Lysholm Knee Score (p=0.037), KOOS-Symptoms (p=0.028), and KOOS-QOL (p=0.029) (Figure 1). There were no significant differences in demographics between those with and without medial tibia BMEP. There was no significant difference in PRO scores when comparing those with medial femoral condyle BMEP, cartilage quality in any of the knee regions, or presence of meniscus extrusion. Conclusions: We observed that patients with medial tibia BMEP at baseline on their preoperative MRI had inferior PROMs for Lysholm Knee Score, KOOS Symptoms, and KOOS-QOL after PMMR repair compared to those without medial tibia BMEP. The presence of medial tibial BMEP could represent underlying damage that may not be evident in the cartilage itself, given most medial tibia cartilage was normal on MRI and cartilage quality was not a significant discriminant for outcomes. The presence of meniscus extrusion did not correlate with patient outcomes either. Limitations of this study include only VAS-Pain score as a pre-operative measure. Multivariate analysis was not possible due to the sample size available. A strength of this study is the mean follow up of approximately 5 years and preoperative MRI data on all patients. Posterior medial meniscus root tears are generally treated with surgical repair due to increased risk of progression to arthritis, however, some patients continue to have symptoms after repair. Our results show that patients with medial tibial BMEP on their preoperative MRI have worse PRO scores after posterior medial meniscus root repair, which can help surgeons better counsel patients towards optimal treatment options. In conclusion, the presence of medial tibia BMEP on preoperative MRI was a marker for inferior patient outcomes for Lysholm Knee Score, KOOS Symptoms, and QOL at minimum two years postoperative following posterior medial meniscus root repair. [Table: see text][Table: see text]

Fast magnetic resonance imaging for diagnosing pulmonary tuberculosis in children: the sub-10-min unenhanced scan.

In this study, we aimed to report the feasibility and quality of fast (unenhanced < 10-min duration) magnetic resonance imaging (MRI) for the detection of lymphadenopathy in non-sedated children with suspected tuberculosis (TB). This was a prospective study that involved children (< 13years of age) hospitalised at Red Cross Children's Hospital with suspected pulmonary TB who were referred for a fast MRI of the chest. The limited short-duration MRI protocol included coronal short tau inversion recovery (STIR) and axial diffusion-weighted imaging (DWI) sequences with additional axial STIR and axial and coronal T2 sequences if the patient was compliant. The scan time was capped at 10min and a study was considered successfully completed when DWI and STIR images were obtained in axial planes. MRI quality was recorded as 'acceptable quality'; 'poor quality, but readable'; and 'non-diagnostic'. Of the 192 fast MRI protocol scans, 166 (86%) were successfully completed within the 10-min allotted scan period. There was no age or sex difference between successful and unsuccessful studies. The mean duration of successful scans was 6.5min (standard deviation = 1.5min, range = 4-10min). Fast (sub-10-min scan) MRI is feasible for diagnosis of lymphadenopathy in non-sedated children in the setting of suspected TB, including those below 6years of age.

Feasibility of accelerated non-contrast-enhanced whole-heart bSSFP coronary MR angiography by deep learning-constrained compressed sensing.

To examine a compressed sensing artificial intelligence (CSAI) framework to accelerate image acquisition in non-contrast-enhanced whole-heart bSSFP coronary magnetic resonance (MR) angiography. Thirty healthy volunteers and 20 patients with suspected coronary artery disease (CAD) scheduled for coronary computed tomography angiography (CCTA) were enrolled. Non-contrast-enhanced coronary MR angiography was performed with CSAI, compressed sensing (CS), and sensitivity encoding (SENSE) methods in healthy participants and with CSAI in patients. Acquisition time, subjective image quality score, and objective image quality measurement (blood pool homogeneity, signal-to-noise ratio [SNR], and contrast-to-noise ratio [CNR]) were compared among the three protocols. The diagnostic performance of CASI coronary MR angiography for predicting significant stenosis (≥ 50% diameter stenosis) on CCTA was evaluated. The Friedman test was performed to compare the three protocols. Acquisition time was significantly shorter in the CSAI and CS groups than in the SENSE group (10.2 ± 3.2min vs. 10.9 ± 2.9min vs. 13.0 ± 4.1min, p < 0.001). However, the CSAI approach had the highest image quality scores, blood pool homogeneity, mean SNR value, and mean CNR value (all p < 0.001) compared with the CS and SENSE approaches. The sensitivity, specificity, and accuracy of CSAI coronary MR angiography per patient were 87.5% (7/8), 91.7% (11/12), and 90.0% (18/20); those per vessel were 81.8% (9/11), 93.9% (46/49), and 91.7% (55/60); and those per segment were 84.6% (11/13), 98.0% (244/249), and 97.3% (255/262), respectively. CSAI yielded superior image quality within a clinically feasible acquisition time in healthy participants and patients with suspected CAD. The non-invasive and radiation-free CSAI framework could be a promising tool for rapid screening and comprehensive examination of the coronary vasculature in patients with suspected CAD. • This prospective study showed that CSAI enables a reduction in acquisition time by 22% with superior diagnostic image quality compared with the SENSE protocol. • CSAI replaces the wavelet transform with a CNN as a sparsifying transform in the CS algorithm, achieving high coronary MR image quality with reduced noise. • CSAI achieved per-patient sensitivity of 87.5% (7/8) and specificity of 91.7% (11/12) respectively for detecting significant coronary stenosis.

No-Reference Image Quality Assessment of Magnetic Resonance images with multi-level and multi-model representations based on fusion of deep architectures

Accurate quality assessment of Magnetic Resonance (MR) images is essential for effective medical diagnostics, as it impacts the time spent on image acquisition and image interpretation by radiologists. Therefore, this study aims to provide a novel deep learning-based No-Reference (NR) MR Image Quality Assessment (IQA) method for the quality prediction of MR images. Therefore, in this work, an internal fusion of two complementary deep learning architectures is introduced that offers MR-specific quality-aware features. Apart from obtaining multi-level image representations of joint networks, to further enhance the quality prediction performance, features from re-trained single deep learning architectures are also used. Finally, as one of the main findings of this study, the quality assessment is performed by a high-level quality model trained on scores of quality models obtained for layers of the networks. The superiority of the method against the state-of-the-art techniques is verified by using two publicly available MRIQA benchmarks containing MR images and subjective scores provided by a large number of radiologists. As reported, the method offers superior IQA of MR images as the obtained scores are highly correlated with subjective opinions of medical specialists. It is characterized by the weighted average values of the Spearman Rank-order Correlation Coefficient, Kendall Rank-order Correlation Coefficient, and Pearson Linear Correlation Coefficient of 0.8754, 0.7185, and 0.9062, respectively.